Written evidence from John Haskey FAcSS AKC, Associate Fellow, University of Oxford [HAB0375]


John Haskey held a Government statistician post for over 15 years as head of the Family Demography Unit in the Office for National Statistics and was responsible for compiling, estimating, analysing, interpreting and publishing official statistics on marriage, divorce, cohabitation and family statistics generally. He has published analyses on all these subjects since the mid-1980s.  From 2000, he was a Visiting Senior Research Fellow at the Department of Social Policy and Intervention, University of Oxford, and is currently an Associate Fellow at the Department.

John received an invitation to submit evidence to the inquiry.

Introduction and considerations

Statistics on cohabitation are fewer than on many subjects such as those on marriage and divorce, primarily because there is no legal obligation to register a cohabitation, whereas the other two events do require legal recording. Consequently, information on cohabitation has to be collected via statistical surveysor, since 1991, from the census. In an interview survey, the question has to be simple, straightforward and readily understandable by the respondent; currently the standard wording is: “Are you living together as a couple?” This direct question does not counter any refinements - which a more detailed definition might demand. This question asks whether the respondent is currently cohabiting. As such, the future outcome of relationship is unknown at the time of the interview, the possibilities are marriage, separation, or continuing to live together. In addition, the respondent can be asked about past completed spells of cohabitation, and how they ended. Furthermore if the cohabitation ended in the partners marrying, then the cohabitation is termed ‘premarital cohabitation’.

Asking questions on cohabitation involves greater uncertainties than with those on marriage, quite apart from the uncertainty arising from deriving estimates from statistical samples. Marriage can be assumed to have started at the date of marriage (although given that the majority of couples premaritally cohabit, the start may be taken as the start of the premarital cohabitation). The starting date of cohabitation – whether pre-marital or otherwise - can be uncertain insofar as it might be the start of sexual relations, or the date of moving in to the same accommodation; the two do not necessarily coincide[1]. Others may consider the date they made a commitment to each other as a couple marked the start of the cohabitation. Partners in the same couple have been known to give different answers as to the start of their cohabiting relationship. Similarly the date of ending of a cohabitation is not always clear-cut, since the ending of the relationship can happen before accommodation-sharing finishes.

Quite apart from these considerations, accurate recall can be difficult. Consequently the duration of the cohabitation can be uncertain, whether the respondent is asked it directly, or asked the start and finish dates, from which the duration can be derived. Whilst these uncertainties apply when asking questions in a statistical survey, they may not be so problematic for a particular cohabiting couple whose situation is being considered in detail, although the possibility remains that the partners do not agree on the duration of their relationship. 

In addition, the partners in a few couples may regard themselves as a couple living together, but not having a sexual relationship.

All these considerations apply to cohabiting same-sex couples as well as to opposite sex couples. Attention will be focussed on opposite sex cohabiting couples.


Trends in cohabitation

There was considerable stigma attached to cohabitation in earlier decades which posed a challenge to measuring it accurately. In 1976, cohabitation was one aspect, but not the main subject of a survey whose purpose was to gauge the influence of a number of factors influencing family formation, and was addressed only to women. The women were asked: “Are you living as married?” and the woman’s legal marital status checked to see whether she was legally married or not. At that time, cohabitants often described themselves as “as married”, although there was a widespread belief that ‘common law marriage’ was equivalent to marriage, and enjoying the same rights – a misconception that persists to the present day.

The prevalence of cohabitation in Great Britain was probably very low during the 1960s; estimated[2] to have been less than 1 per cent amongst adults aged under 50. Even in 1976 from the Family Formation Survey[3] just mentioned, the proportion of women cohabiting was found to be only a few percent.

Table 1 gives a profile of the main kinds of families, including cohabiting couple families, for each of the stated years, and with successive ‘snapshots’ of the percentage[4] of each kind of family each year.


Overall, it will be seen that cohabiting couple families have quadrupled in relative size as a proportion of all families from 1986 to 2020, with cohabiting couples with no children consistently forming roughly one half of all cohabiting couple families. Each of the three categories of cohabiting couple families has grown in relative size, approximately trebling or quadrupling over the period concerned. In contrast, married couple families have formed a steadily decreasing proportion of all families, with the proportion of each of the three kinds of


married couple family declining by around one quarter from 1986 to 2020. Whereas married couples with dependent children formed one half of all families in 1986, by 2020 they accounted for a little over one third. Lone parent families generally increased as a proportion of all families.


It is estimated that there were 2.0 million cohabiting couples in England and Wales in 2002, which rose to 3.2 million in 2019, an increase of 60 per cent. However the overall population increased between 2002 and 2019, and this increase can be taken into account by considering the proportion of the adult population who were cohabiting. Because the number of men in the population is slightly smaller than the number of women, the proportion of men cohabiting is slightly larger than that of women. Figure 1 shows the increase in the proportions of men and women cohabiting.

For simplicity, attention will be restricted to women cohabiting as the patterns are very similar between men and women.

The proportions cohabiting increased in every age group between 2002 and 2019 for both men and women, and a linear increase for each age group i.e. a steady constant upward gradient. However, the magnitude of the increase in the proportion of women cohabiting has varied according to age group, as Figure 2 shows.

In general, the smaller the proportion of women cohabiting in 2002, the larger the subsequent increase in the proportion. So, although the proportion of women aged 30-34 who were cohabiting was the largest of all age groups in both 2002 and 2019, the associated percentage increase in the proportion cohabiting between 2002 and 2019, 37 per cent, was the second smallest percentage increase of all age groups of women. The smallest percentage increase, 22 per cent, occurred for the age group 16-29, again indicating that in general, the larger the proportion in 2002, the smaller the increase in the proportion cohabiting. In contrast, the proportion of women aged 50-54 who were cohabiting doubled (100 per cent increase), and quadrupled (330 per cent increase) for those aged 60-64. However, it should be borne in mind that the potential for a relatively large increase is greater for those whose proportion cohabiting was initially comparatively small. Overall, these results indicate that the age profile of cohabiting women has become older.

The fact that the proportionate increase in cohabitation has varied considerably according to age is partly explained by the different profiles of legal marital statuses. Figure 3 illustrates

that the youngest age group consist almost entirely of those who have never had a formalised legal relationship, whether marriage or a civil partnership; that is, they are predominantly single. In contrast, the older the age group, the larger the proportion cohabiting who had a previous divorce or a dissolution of a civil partnership. It is notable that, amongst the age groups, the larger the proportion who had divorced or who had dissolved their civil partnership, the larger the increase in the proportion cohabiting between 2002 and 2019.



Numbers cohabiting by age and sex

As well as considering the proportions cohabiting, and how much they have increased, it is appropriate to keep in mind the actual numbers cohabiting in the different age groups. Table 2 shows these estimated numbers for 2019. It may be seen that the majority of women cohabiting are aged under 35; they represent 51 per cent of all women cohabitants, whereas the majority of men cohabitants are aged under 40, accounting for 57 per cent of all men cohabitants. Relatively few cohabitants are aged 60 or more; 11 per cent of all men cohabitants and 9 per cent of all women cohabitants.


Further analyses of cohabitation from several studies

Some key results are given below, drawn from several sources. Some of the findings have been given in generalised form, whilst others have quoted the (prevailing) published proportions. Although the actual numbers and proportions are most likely to have changed, the trends and relationships are likely to have continued and persisted. A number of results have been extracted from a comprehensive analysis[5] of cohabitation in Britain which was published in 2011, based on retrospective histories from the General Household Survey 1979-2007. Further results have been drawn from some other sources[6],[7].

(Note: In the results from the comprehensive analysis mentioned above, the comparisons over time typically refer to marriages taking place in – or cohabitations started in - 1980-84 (“early 1980s”) and 2004-2007 (“mid-2000s”), unless otherwise stated.)





























Attitudes to marrying one’s partner

With the gradual disappearance of stigma attached to many informal relationships, attitudes have undergone a substantial change since the 1960s. Attitudes, intentions, and aspirations understandably vary according to one’s age, marital status and living arrangements. Figure 4 illustrates the expressed likelihood in 2019 of cohabiting respondents marrying their partner in the future. The proportions for the three possible answers are shown for each age group (the three proportions totalling to 100 per cent).  At the youngest ages, the majority say that it is likely or very likely they will marry their partner, with the peak occurring in the early twenties when the proportion is two thirds. This proportion quickly declines with age, and by the late thirties, the proportion expressing doubt (“not sure”) about the likelihood of marrying their

partner has grown to almost one half, being more than double that of saying it is likely. And then, from the late forties and older ages, the predominant reply is that they are unlikely or very unlikely to marry their partner. Paradoxically, despite these results, marriage rates for older men and women have increased over the last ten years, with the largest increases being recorded for those aged 65 and over[10]. Also, marriage rates have fallen for the younger ages, particularly so for teenagers. However, Figure 4 gives a ‘snapshot’ of attitudes in 2019, not trends over time.

The same respondents were then asked their views of the future of their relationship, and although the question was less definite, the alternative answers offered were generally more definite – see Figure 5. Although the categories cannot be directly compared between the two graphs, the pattern with age is much more distinct in the second graph. The proportions definitely planning to marry are generally smaller than those likely or very likely to marry (as might be expected), and fall away more rapidly after the early thirties. For those aged 50 and over, the proportions who say they were planning to marry are exceedingly small. Unlikely or very unlikely to marry one’s partner might roughly equate to a probability of continued living together, although the proportions for the latter are consistently slightly larger than those for the former in each age group, especially for those in their forties and fifties.

There may be a tendency to overstate the possibility of marriage when likelihood terms are used rather than terms of probability, which may sound more committed, and final. For those aged 40 and over, the majority state they will probably keep on living together rather than marrying, and these proportions are larger than the corresponding ones stating they are unlikely or very unlikely to marry their partner. These findings accord with the growth in the proportions cohabiting; the largest increases over the past decade have occurred for those aged 40 and over. Those who are oldest and not married may have been with their partner for longer than others, which predisposes them to continue further without marrying.


Attitudes of respondents living with a partner concerning the advantages and disadvantages of living as a couple

It is of interest to consider the attitudes which cohabiting partners hold, and in particular those which might have influenced them when starting to live together, rather than marrying.

One question in Understanding Society[11] was addressed to couples living together[12], and they were first asked: “Do you think there are any advantages in living as a couple, rather than being married?” Those who answered “Yes” were then asked: “What do you think are the advantages of living as a couple?” A total of 11 possible answers were presented to them, and they were asked to select all that applied. (The 11 possible answers, complete in wording as presented, are given in Table 3.) The answers can be divided roughly into three groups: ones expressing caution or precaution (5 in number, indicated by C in Table 3), another group mentioning positive or financial advantages (4 in number, indicated by P), and the third group being other or neutral (2 in number, indicated by O).


Table 3  Possible answers to the question: What do you think are advantages of living together as a couple, rather than marrying?


1                     Trial marriage, see if we’re compatible before committing to marriage, C

2                     No legal ties so easier to split up if doesn't work out, C

3                     Improves relationship – more respect, less taking partner for granted, get on  better, less arguments etc. ..., P             

4                     A lot of marriages fail, C             

5                     Personal independence-can keep own privacy, not ready for marriage, etc., C

6                     It does not have to be a permanent commitment, C             

7                     Tax advantages, P             

8                     Saves on the cost of a wedding, P             

9                     Companionship and someone to share things with, P             

10                No particular advantage, just prefer cohabitation, O

11                Other, O             



Overall, the proportions of men and women who replied positively to the first question – whether they thought there were advantages in living together - were very similar; both 43 per cent. (These were the proportions answering Yes out of all who answered either Yes or No.)  Furthermore, there was little difference in the proportions replying positively between men and women in each age group, and, in addition, the average numbers of answers, i.e. different identified advantages, given by men and women were very similar: 2.9 and 2.8, respectively. 

From the total of 11 possible answers of the advantages of living together, rather than marrying, the most frequently mentioned answer by men was “improves the relationship, more respect, etc.”, at 16 per cent of all men’s mentions, whereas that for women was “companionship and someone to share things with” at 16 per cent of all women’s mentions. The second most mentioned answer for men was “companionship, etc.” at 16 per cent, and that for women was “No legal ties, so easier to split up if doesn't work out” at 15 per cent. It is interesting that high up on the list for women is, effectively, living together being a way of avoiding divorce. In fact, the same answer also appears high on the list for men; it ranked third most important, accounting for 13 per cent of all men’s mentions. A Dutch study[13] of 40 respondents, interviewed in 7 focus groups, indicated a general concern about divorce which coloured all their views on cohabitation and marriage. For the higher educated, divorce was seen as a potential personal failure, while less educated women saw cohabitation as carrying less financial risk than marriage, which could involve debt-sharing. Respondents generally accepted cohabitation as a strategy to test compatibility.

The most frequently mentioned answers in the different age groups give some understandable results. For those in their twenties, the answer with the most frequent mentions was “trial marriage, to see if we are compatible before committing to marriage”; whilst those in their thirties was “saves on the cost of a wedding”. For those in their forties, fifties and fifties and older, the most frequent mentions were each for “companionship and someone to share things with”. These answers obviously reflect on the respondents’ stage of life, and what most interested, concerned, or applied to them in their current situation.

There are several of the possible 11 answers which are of particular interest: two and possibly three reflect awareness of divorce, and a desire to avoid it; whereas another two reflect a reluctance to be committed, or fully committed, to one’s partner. Mentions of these answers can informatively be analysed by age group, since as recognised immediately above, different emphases are apparent at different stages in one’s relationship lifetime.  












Figure 6  For each age-group, the percentage proportion of the total



mentions of all the answers made by that age group which were accounted


for by the mentions of the specified answer, 2019, England and Wales





























































The top graph in Figure 6 gives the proportions of mentions for “trial marriage, etc.”, and for “companionship, etc.” Although the former is a “caution-type” answer, and the latter a “positive-type” answer, both answers involve either exploring the potential for companionship, or else contemplating achieving it. Overall, i.e. over all age groups, mention of “trial marriage, etc.” and “companionship, etc.” formed 9 per cent and 16 per cent of all mentions, respectively, so that variations by age about these averages indicate whether they were deemed relatively more, or less, advantageous than the average.

It may be seen from Figure 6 that the proportions of mentions of “trial marriage, etc.” for those in their early twenties[14], late twenties and early thirties are 17, 19 and 19 per cent, respectively, considerably larger than the overall proportion of 9 per cent. Not surprisingly, the respondents in these age groups were the most likely to consider living together with a partner to see if their relationship would work. In contrast, the proportions for each subsequently older age group are smaller than that average, and furthermore, become progressively smaller the older the age. Again, it is not surprising that older respondents living as a couple do not think, either for themselves, or their contemporaries, in terms of a trial marriage, which they probably consider as more appropriate for younger couples. Indeed, these results are consistent with those from Figure 5 which indicated that the vast majority of those older than about 45 said they would probably just keep living together without marrying.

Also shown in the top graph in Figure 6 are the proportions for “companionship, etc.”, and overall, this answer accounted for 16 per cent of all mentions. The pattern of the proportion of mentions with age is completely different from that of “trial marriage, etc.”, with below average mentions for those in the youngest age groups. The proportion generally increases with age, and those aged over 60 were almost twice as likely to mention companionship as those in their twenties. (The proportions of mentions for both “trial marriage, etc.” and “companionship, etc.” are both below their average for those in their late thirties, whose most mentioned answer was “saves on the cost of a wedding”. Possibly this age group is a turning point between identifying with a trial marriage and valuing companionship.)

The lower graph in Figure 6 considers two “cautionary” answers, both of which indicate a reluctance to give up “one’s freedom”; in one, “personal independence, etc.”, the respondent admits they are not ready (or prepared for) marriage, whereas in the other, the respondent values living together as not necessarily involving a long-term commitment (which marriage would imply). Overall, “personal independence, etc.” accounts for 8 per cent of all mentions, whereas “not permanent commitment, etc.” accounts for 5 per cent of all mentions.  The two sets of proportions of mentions are very similar for the age-groups from the late twenties to the late thirties, but for older age groups, above age 50, the proportions of mentions for “personal independence, etc.” tend to be above average, whereas those for “not permanent commitment, etc.” are small and below average. This pattern is understandable; at older ages a permanent commitment may seem inappropriate to a respondent who has been living with his or her partner for years (or if the respondent has been divorced and re-partnered), whilst “personal independence, etc.” may be increasingly valued and experienced by respondents in the older age groups who have their own interests to pursue - and tolerant of a mutual degree of independence - in their relationships.


A parallel introductory question to that on the advantages of living together was then asked of respondents living as a couple: ““Do you think there are any disadvantages in living as a couple, rather than being married?” Overall, 33 per cent of those who responded either “Yes” or “No” replied positively (31 per cent of men respondents, and 35 per cent of women respondents). Those who responded that there were disadvantages were then asked to select as many reasons as applied, from a list of six. The most frequently mentioned reason was that living together provided “no legal status”[15] which accounted for 40 per cent of all mentions. (The second most mentioned reason was “financial insecurity” at 27 per cent.) The least mentioned specific reason was “lack of commitment” which registered 9 per cent of all mentions. (The other possible answers were “effect on children”, 11 per cent of all mentions; “social stigma”, 11 per cent of all mentions; and “other”, 3 per cent of all mentions.)

Attention will be focussed on “no legal status” and “lack of commitment”, which are arguably the factors most likely to affect the stability of cohabiting unions. Overall, men and women ranked “no legal status” equally as the most important disadvantage; 40 per cent of men’s mentions were for this reason, and similarly 40 per cent of women’s. Women mentioned “lack of commitment” proportionately more frequently than men; 10 per cent of all mentions by women, compared with 7 per cent of all mentions by men. The proportions of all mentions of “no legal status” and “lack of commitment” are shown in Figure 7, separately for each age group[16].


In general, the proportion of mentions of “no legal status” increases steadily with age, with the late forties being a turning point; for younger ages the proportions are all below average, and for older age groups the proportions are above average, and increase with age. Legal status, and in particular, not being married, has financial disadvantages if the cohabiting union should end in separation; there can be other financial implications[17] of not being married at older ages, such as pensions and inheritance. Possibly there is greater awareness of these implications amongst older respondents. In addition, some of these older respondents would have previously been married and would have had first-hand experience of losing married status.

In contrast, mentions of “lack of commitment” are most frequent in the early thirties and early forties, after which the proportion generally declines with increasing age. It is unlikely to be coincidental that mentions of “lack of commitment” are relatively most numerous in those age groups, the thirties and forties, for which the divorce rates are highest. Possibly couples living together are aware of these hazards to their married counterparts, and wonder about the vulnerability of their partnerships to the same hazards. Lack of commitment may be associated with being unsure of the permanence of one’s relationship, and in a study[18] of commitment and relationship quality in Sweden and Norway, cohabiting respondents who had no plans to marry were only about one third as likely to be serious about their current union compared with those who were married.


As has been seen, the identified advantages and disadvantages of not marrying included: trial marriage; no legal ties; marriages failing; no permanent commitment; personal independence; and no legal status, are all either directly, or indirectly, associated with recognising or avoiding divorce.


It should be mentioned that the possible answers to the questions on the advantages and disadvantages of living together as a couple were pre-specified (presumably identified from volunteered answers in a pilot survey), and other answers might have been given had the questions been completely open-ended. Given the recent history of opposite sex civil partnership, one might have expected some reasons for living together rather than marrying was because marriage was seen as paternalistic with unequal status between the spouses. However, both sets of possible answers included an ”Other” category which accounted for very small proportions; 1.6 per cent and 2.7 per cent of all mentions of advantages, and disadvantages, respectively. Consequently if this attitude was volunteered as an advantage of living together, it was evidently considerably out-ranked in importance by the specified reasons.


August 2021




[1] John Haskey, Cohabiting couples in Great Britain: accommodation sharing, tenure and property ownership, Population Trends, 103, 2001, pp.26-36.

[2] Murphy, M. The evolution of cohabitation in Britain, 1960-95, Population Studies, 2000, 54, pp.43-56.


[3] Dunnell, Karen, Family Formation 1976: a Survey Carried Out on Behalf of Population Statistics Division 1 of the Office of Population Censuses and Surveys of a Sample of Women (both Single and Ever Married) Aged 16-49 in Great Britain, London: H.M. Stationery Off., 1979. Print.

[4] The percentages for the earlier years in Table 1 were published at the time, and, subsequently, technical revisions could imply that the percentages might need slight revision.

[5] Beaujouan and Ni Bhrolchain, Cohabitation and marriage in Britain since the 1970s, Population Trends, 145, 2011, pp. 35-59.

[6] John Haskey, Cohabitational and marital histories of adults in Great Britain, Population Trends,96, 1999,

pp. 13-23.

[7] Office for National Statistics, Population estimates by marital status and living arrangements, England and Wales: 2019.

[8] Kiernan and Smith, Unmarried parenthood: new insights from the Millennium Cohort Study, Population Trends, 114, 2003, pp. 26-33.


[9] See Footnote 2

[10] ONS Statistical Bulletin: Marriages in England and Wales: 2017

[11] A longitudinal survey covering all household members

[12] Of the cohabiting couples living together, 2.3 per cent were same sex couples

[13] N. Hiekel and R. Keizer Risk-avoidance or utmost commitment? Dutch focus group research on cohabitation and marriage, Demographic Research 32, 2015, pp.311-340. https://doi.org/10.4054/DemRes.2015.32.10

Hiekel, N., & Keizer, R. (2015). Risk-avoidance or utmost commitment? Dutch focus group research on cohabitation and marriage. Demographic Research, 32, 311-340

[14] As an example, amongst 20 to 24 year olds, 76 mentions were made of “trial marriage, etc.” as an advantage of living as a couple, whereas a total of 208 mentions were made by this age group of all the different answers. As a result, for the age group 20 to 24 years old, the proportion of all mentions accounted for by “trial marriage, etc.” was 100 * ( 76/208) or 17.3 per cent. (The total of all the percentages over the different possible answers, eleven in number, of course, is 100 per cent, and this applies to each age group.)

[15] Whilst it is a mistaken belief that cohabiting partners enjoy similar rights to married couples, it is also a myth that cohabitation has no legal significance.

[16] As an example, amongst 20 to 24 year olds, 22 mentions were made of “no legal status” as a disadvantage of living as a couple, whereas a total of 72 mentions were made by this age group of the different answers. As a result, for the age group 20 to 24 years old, the proportion of all mentions accounted for by “no legal status” was 100*( 22/72) or 30.6 per cent. Comparing this percentage between different age groups indicates the relative importance the different age groups attach to this particular disadvantage of living in a couple. (The total of all the percentages over the different possible answers, six in number, of course, is 100 per cent, and this applies in each age group.)

[17] Women were more likely to mention “financial insecurity” as a disadvantage than men: 29 per cent of all mentions by women, and 25 per cent of all mentions by men.

[18]Kenneth Aarskaug Wiik, Eva Bernhardt, Turid Noack, A Study of Commitment and Relationship Quality in Sweden and Norway, 2009, Journal of Marriage and Family, Vol. 71, Issue 3, pp. 465-477.https://ezproxy-prd.bodleian.ox.ac.uk:2102/10.1111/j.1741-3737.2009.00613.x